remote proc kick failed when rpmsg is launched by U-Boot

[hfranco]

Hello,

When the rpmsg_lite_str_echo demo from MCUXpresso is launched by the remoteproc framework, the RPMSG

demo works fine, the channel is created and I can send messages between the two cores using /dev/ttyRPMSG30 endpoint.

However, when it is launched by U-Boot bootaux command, when the imx_rpmsg_tty demo driver is modprobed,

the following  error happens:

    root@colibri-imx8x-06858909:~# modprobe imx_rpmsg_tty

    [ 23.060068] imx_rpmsg_tty virtio1.rpmsg-openamp-demo-channel.-1.30: new channel: 0x400 -> 0x1e!

    [ 23.069286] Install rpmsg tty driver!

    [ 23.174246] imx-rproc imx8x-cm4: imx_rproc_kick: failed (3, err:-62)

    root@colibri-imx8x-06858909:~# echo "teste" >/dev/ttyRPMSG30

    [ 52.578255] imx-rproc imx8x-cm4: imx_rproc_kick: failed (3, err:-62)

    [ 52.686242] imx-rproc imx8x-cm4: imx_rproc_kick: failed (3, err:-62)

The code returns TIMER EXPIRED (-62) at (imx_rproc.c)

      /*

       * Send the index of the triggered virtqueue as the mu payload.

       * Let remote processor know which virtqueue is used.

       */

      mmsg = vqid << 16;

      err = mbox_send_message(priv->tx_ch, (void *)&mmsg);

      if (err < 0)

            dev_err(priv->dev, "%s: failed (%d, err:%d)\n",

                  __func__, vqid, err);

where (mailbox.c):

      if (chan->cl->tx_block) {

            unsigned long wait;

            int ret;

            if (!chan->cl->tx_tout) /* wait forever */

                  wait = msecs_to_jiffies(3600000);

            else

                  wait = msecs_to_jiffies(chan->cl->tx_tout);

            ret = wait_for_completion_timeout(&chan->tx_complete, wait);

            if (ret == 0) {

                  t = -ETIME;

                  tx_tick(chan, t);

            }

      }

I inserted some print debugging in loop on the cortex-m side and looks like the cortex-m stops running

when the kick failed happens, indicating that the cortex-m might have crashed. Do you know why this is happening?

Any suggestions to debug this issue?

Best regards,
Hiago.

[Manuel_Salas]

Hello @hfranco 

I hope you are doing very well.

Could you please clarify the processor you are using, BSP version, SDK version and if possible if you are using an EVK from NXP?

Best regards,

Salas.

[hfranco]

Hello @Manuel_Salas,

I am using the Colibri i.IMX8QXP from Toradex. BSP version is based on NXP 6.6 kernel, from linux-imx. SDK version is SDK_2_9_0_MIMX8UX5xxxFZ.

For further information, setting 'pd_ignore_unused' as kernel argument fixes the issue.

Regards,
Hiago.